QMIX implemented in TensorFlow 2

Note: Currently, we are only experimenting with Two State Game.

How to run QMIX in Two State Game

First, install packages.

pip install -r requirements.txt

Run main.py

python main.py

Then, outputs episode reward and qmix's loss history graph.

Reference

• QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning